A PN junction reversal problem

The model I created is to add a P-type region at the top of the uniformly distributed N-type region, apply voltage to the bottom boundary, the top boundary voltage is 0, and the steady-state solver does not converge.